Transformer insulation is the single most critical component determining the life and reliability of a power transformer. Over 60% of transformer failures are linked to insulation breakdown. Once it fails, you’re looking at months of downtime and repair costs that can reach hundreds of thousands of dollars.
The good news: insulation problems don’t happen overnight. They develop over years, and the right testing program will catch them early. This guide covers every test method — from a simple megger check during commissioning to advanced partial discharge measurements on 220 kV units — with the exact standard requirements from IEC 60076-3 and IEEE C57.152.
Table of Contents
Why Transformer Insulation Fails
Transformer insulation degrades through a combination of thermal, electrical, chemical, and mechanical stresses. Understanding these helps you test at the right times and interpret results correctly.
Thermal aging — This is the primary failure mechanism. Cellulose insulation (Kraft paper, pressboard) breaks down chemically when exposed to heat. Every 6°C rise above the rated hot spot temperature roughly halves the insulation life. Overloaded transformers age their insulation dramatically faster than those running at rated load.
Moisture ingress — Water is the enemy of transformer insulation. Even small amounts of moisture — as low as 2–3% by weight in the paper — dramatically reduce both dielectric strength and the rate of thermal aging. Moisture enters through failed gaskets, breathing cycles, and degradation of the cellulose itself (aging produces water as a byproduct).
Electrical stress — Repeated voltage spikes from lightning strikes, switching operations, and load rejection events stress the insulation at its weakest points. Over time, this leads to partial discharge activity — tiny electrical breakdowns inside the insulation that erode it from within.
Contamination — Particles, sludge, and dissolved decay products in the oil reduce its insulating properties. Sludge deposits on windings act as thermal blankets, trapping heat and accelerating aging.
Mechanical stress — Through-fault currents from external short circuits generate enormous mechanical forces in the windings. Repeated faults can loosen windings, displace insulation, and create new paths for electrical failure.
Types of Insulation in Transformers
Most power transformers use an oil-paper insulation system consisting of:
Kraft paper — Wrapped around individual conductors. This is the primary turn-to-turn insulation. Its condition largely determines the transformer’s remaining life.
Pressboard — Structural insulation components that maintain spacing between windings, between windings and core, and between windings and tank. Thicker and more rigid than Kraft paper.
Crepe paper — Flexible insulation used on curved surfaces and conductor bends.
Transformer oil (mineral oil) — Fills all spaces, providing both insulation and cooling. The oil’s dielectric strength depends on its purity, moisture content, and dissolved gas levels.
Dry-type transformers use different insulation systems — typically cast resin (Class F or H) or VPI (vacuum pressure impregnated) windings. The testing approach differs, and I’ll note where it matters.
Insulation Resistance (IR) Testing
The basics
IR testing on transformers follows the same Ohm’s law principle as motor testing: apply DC voltage, measure leakage current, calculate resistance. But transformer testing has some important differences.
Test connections
For a two-winding transformer, you make three separate measurements:
- HV winding to ground — with LV winding shorted and grounded
- LV winding to ground — with HV winding shorted and grounded
- HV winding to LV winding — with both windings’ opposite terminals grounded
For three-winding transformers, add combinations for each winding pair.
Always short all terminals of each winding together before connecting to the megger. This ensures you’re measuring the insulation between windings and ground, not between individual turns.
Test voltage selection
The test voltage depends on the transformer’s voltage class:
| Transformer Winding Voltage | IR Test Voltage (DC) |
|---|---|
| Up to 1 kV | 1,000V |
| 1 kV – 5 kV | 2,500V |
| 5 kV – 15 kV | 5,000V |
| Above 15 kV | 5,000V – 10,000V |
For HV machinery electrical equipment, IEC 60204-11 Clause 19.3 specifies: test voltage equals the rated voltage of the equipment or 5 kV, whichever is lower, with a minimum IR of 1 MΩ.
Minimum values
There’s no single universal minimum for transformer IR. The actual value depends on the transformer’s size, voltage class, temperature, and the type of insulation.
General guidelines for oil-filled transformers:
| Condition | Typical IR Range |
|---|---|
| New transformer, clean dry oil | 500 MΩ – 10 GΩ+ |
| In-service, good condition | 100 MΩ – 1 GΩ |
| Aging but acceptable | 50 MΩ – 100 MΩ |
| Needs investigation | Below 50 MΩ |
Critical rule: the trend matters more than the absolute value. A transformer that reads 80 MΩ consistently for 5 years is fine. A transformer that dropped from 300 MΩ to 80 MΩ in one year needs investigation.
Temperature correction
Transformer IR readings are heavily temperature-dependent. The resistance roughly halves for every 10°C rise. IEEE C57.152-2013 recommends correcting to 20°C for comparison.
Correction formula: R₂₀ = R_measured × C
Where C is the temperature correction factor from the standard’s tables. As a simplified approximation: C = 2^((T − 20) / 10), where T is the oil temperature in °C.
Polarization Index (PI) for Transformers
How it differs from motor PI
The PI works the same way — 10-minute reading divided by 1-minute reading. But transformer interpretation is different because transformer oil has a polarization index close to 1.0 by nature. This means transformer PI values are typically lower than motor values.
Interpretation per IEEE C57.152-2013
IEEE C57.152-2013, Clause 7.2.13.4 provides the following PI interpretation for transformers:
| PI Value | Transformer Insulation Condition |
|---|---|
| Below 1.0 | Dangerous — do not energize |
| 1.0 – 1.1 | Poor — serious degradation likely |
| 1.1 – 1.25 | Questionable — investigate further |
| 1.25 – 2.0 | Fair — acceptable but monitor closely |
| Above 2.0 | Good — insulation in healthy condition |
Compare this to IEEE 43-2013 for motors: a PI of 1.5 on a motor (Class A) is the bare minimum. On a transformer, 1.5 is “fair” to “good.” Don’t apply motor PI criteria to transformers — you’ll condemn perfectly healthy units.
When PI is useful on transformers
The PI test is most valuable for dry-type transformers and for the solid insulation components of oil-filled transformers. For transformers with new, high-resistivity oil, the PI will naturally be low because the oil dominates the measurement. This doesn’t indicate a problem.
PI is particularly useful when the oil has been drained for inspection — you can test the solid insulation directly without the oil masking the reading.
Power Factor / Tan Delta Testing
What it measures
Power factor (also called dissipation factor or tan delta) testing measures the dielectric losses within the insulation. It applies an AC voltage and measures the ratio of resistive current (loss current) to capacitive current.
Tan δ = IR / IC
Good insulation is almost purely capacitive — the tan delta is very low. As insulation ages, absorbs moisture, or becomes contaminated, the resistive component increases and the tan delta rises.
Typical values
| Insulation Condition | Tan Delta (%) |
|---|---|
| New, clean, dry oil-paper insulation | < 0.5% |
| Good in-service condition | 0.5% – 1.0% |
| Aging or contaminated | 1.0% – 2.0% |
| Needs investigation | > 2.0% |
For current transformers specifically, from my commissioning experience on 220 kV and 66 kV systems, acceptable tan delta values should be below 0.5% (5 × 10⁻³). Any elevation above this signals partial discharge risk or contamination.
When to use it
Tan delta testing complements IR testing. IR testing measures the DC resistance of the insulation. Tan delta testing measures losses under AC conditions, which better represents the insulation’s behavior during normal operation.
I typically run tan delta tests on all transformers above 66 kV during commissioning and as part of major maintenance outages.
Dissolved Gas Analysis (DGA)
The most powerful diagnostic tool
DGA is the single most important condition monitoring tool for oil-filled transformers. It doesn’t test insulation resistance directly, but it detects the products of insulation breakdown.
When insulation degrades — whether from overheating, arcing, partial discharge, or chemical deterioration — it produces gases that dissolve in the transformer oil. By analyzing these gases, you can identify the type and severity of the fault.
Key gases and what they indicate
| Gas | Primary Source |
|---|---|
| Hydrogen (H₂) | Partial discharge, low-energy arcing |
| Methane (CH₄) | Low-temperature thermal fault (150–300°C) |
| Ethane (C₂H₆) | Low-temperature thermal fault (150–300°C) |
| Ethylene (C₂H₄) | High-temperature thermal fault (300–700°C) |
| Acetylene (C₂H₂) | Arcing, high-energy discharge (>700°C) |
| Carbon monoxide (CO) | Cellulose (paper) degradation |
| Carbon dioxide (CO₂) | Cellulose degradation |
CO and CO₂ are particularly important because they specifically indicate degradation of the cellulose insulation (Kraft paper, pressboard). A rising CO/CO₂ ratio over time is one of the strongest indicators of insulation aging.
How DGA relates to insulation testing
DGA catches problems that electrical tests miss. A transformer can have acceptable IR and PI values while internal partial discharge is slowly destroying the insulation from the inside. DGA will detect the gases produced by that partial discharge activity.
The best maintenance programs combine electrical tests (IR, PI, tan delta) with DGA for a complete picture.
Partial Discharge (PD) Measurement
What it is
Partial discharge testing detects tiny electrical breakdowns occurring inside the insulation — discharges that don’t fully bridge the gap between conductors but erode the insulation progressively.
PD requirements per IEC 60076-3-2013
IEC 60076-3-2013 specifies the induced voltage test with partial discharge measurement (IVPD) for power transformers. The standard defines strict acceptance criteria (Clause 11.3.5):
Test acceptance criteria:
- PD levels during the one-hour measurement period must not exceed 250 pC
- PD levels must not show any rising trend during the test
- No sudden sustained increase during the last 20 minutes
- PD levels must not increase by more than 50 pC during the one-hour period
- PD at 1.2 × Ur/√3 after the one-hour period must not exceed 100 pC
- Background PD level must not exceed 50 pC at both the beginning and end of the test (100 pC for shunt reactors)
When PD testing is required
Per IEC 60076-3-2013, Table 1:
| Transformer Category | IVPD Test Status |
|---|---|
| Um ≤ 72.5 kV | Special test (if specified by purchaser) |
| 72.5 kV < Um ≤ 170 kV | Routine test |
| Um > 170 kV | Routine test |
For transformers above 72.5 kV, partial discharge measurement during the induced voltage test is a routine requirement — not optional.
Factory Dielectric Tests per IEC 60076-3-2013
Overview of required tests
IEC 60076-3-2013 defines a comprehensive set of dielectric tests. The requirements depend on the highest voltage for equipment (Um) of the highest voltage winding.
Required tests by transformer category (Table 1):
| Test | Um ≤ 72.5 kV | 72.5 < Um ≤ 170 kV | Um > 170 kV |
|---|---|---|---|
| Full wave lightning impulse (LI) | Type test | Routine | Part of LIC |
| Chopped wave impulse (LIC) | Special | Special | Routine |
| Switching impulse (SI) | N/A | Special | Routine |
| Applied voltage (AV) | Routine | Routine | Routine |
| Induced voltage withstand (IVW) | Routine | Routine | N/A |
| Induced voltage with PD (IVPD) | Special | Routine | Routine |
| Auxiliary wiring (AuxW) | Routine | Routine | Routine |
Test voltage levels (Table 2 — selected values)
IEC 60076-3-2013, Table 2 provides the standard test voltage levels. Here are the most commonly encountered values:
| Um (kV) | Lightning Impulse LI (kV peak) | Applied Voltage AV (kV rms) |
|---|---|---|
| <1.1 | — | 3 |
| 3.6 | 20 / 40 | 10 |
| 7.2 | 60 / 75 | 20 |
| 12 | 75 / 95 | 28 |
| 24 | 125 / 145 | 50 |
| 36 | 170 / 200 | 70 |
| 72.5 | 325 / 350 | 140 |
| 145 | 550 / 650 | 230 / 275 |
| 245 | 850 / 950 / 1050 | 360 / 395 / 460 |
| 420 | 1175 / 1300 / 1425 | 510 / 570 / 630 |
The lower values in each row represent the minimum standard level. Higher values provide greater margin.
Test sequence (Clause 7.2.3)
The standard specifies this test order:
- Lightning impulse tests (LI, LIC, LIN, LIMT)
- Switching impulse (SI)
- Applied voltage test (AV)
- Line terminal AC withstand (LTAC)
- Induced voltage withstand (IVW)
- Induced voltage with PD measurement (IVPD) — always last
Testing after service (Clause 8)
For transformers repaired to restore functionality after breakdown:
- New parts: tested at 100% of original test voltage
- Used parts verified for continued use: 80%–100% of original test voltage
- IVPD test: always at 100% of original test voltage
For complete rewinds intended to restore “as new” condition: all routine tests at 100%.
Auxiliary Wiring Insulation Tests
IEC 60076-3-2013, Clause 9 specifies separate insulation tests for transformer auxiliary wiring — the control circuits, instrumentation, and CT secondary wiring that aren’t part of the main power windings.
Requirements from the standard
| Wiring Type | Test | Duration | Pass Criteria |
|---|---|---|---|
| Auxiliary power and control wiring | 2 kV AC to earth | 1 minute | No voltage collapse |
| CT secondary wiring | 2.5 kV AC to earth | 1 minute | No voltage collapse |
| CT wiring (knee point >2 kV) | 4 kV AC to earth | 1 minute | No voltage collapse |
| Site wiring check (after transport) | 1 kV DC IR measurement | — | ≥ 1 MΩ |
Important exclusions from the test circuit:
- All solid-state and microprocessor-based devices — disconnect before testing
- All three-phase undervoltage relays — remove from circuit
- Withdrawable type devices — remove from circuit
- Motors and apparatus — tested per their own IEC standard (which may have lower requirements)
Practical note from field experience: The standard notes that it is normal practice to check all auxiliary wiring at 1 kV DC for 1 minute on site before energization. I follow this practice on every transformer commissioning. It catches wiring errors, damaged cables from transport, and contamination from construction work.
Current Transformer Insulation Testing
CTs integrated into power transformers or installed separately at substations need their own insulation testing. Based on my 12 years commissioning CTs on 66 kV and 220 kV systems, here are the specific test voltages and procedures I use:
IR test connections and voltages
| Test Connection | Test Voltage (DC) | Duration | Minimum IR |
|---|---|---|---|
| Primary winding to ground | 5,000V | 60 seconds | >100 MΩ |
| Primary winding to secondary winding | 2,500V | 60 seconds | >100 MΩ |
| Between secondary cores (multi-core CTs) | 1,000V | 60 seconds | >100 MΩ |
| Secondary winding to ground | 1,000V | 60 seconds | >100 MΩ |
On new HV current transformers, I typically see readings well into the GΩ range. Anything below 100 MΩ warrants investigation.
Additional CT tests
Beyond insulation resistance, a complete CT commissioning includes:
Tan delta test — Mandatory for HV CTs (220 kV, 500 kV). Acceptable values below 0.5% (5 × 10⁻³).
Winding resistance — Measures the ohmic resistance of primary and secondary windings. Secondary winding values should be close to manufacturer data (generally <0.1 Ω).
Polarity test — Verifies correct orientation of primary and secondary windings. Incorrect polarity can cause protection relay maloperation.
Turns ratio test — Compares measured ratio to nameplate value. Deviation >0.5% indicates a winding problem.
Excitation curve and knee point voltage — Traces the magnetization characteristic. The knee point voltage must match specifications — critical for protection-class CTs to avoid saturation during fault conditions.
Field Testing After Installation
Before first energization
Every transformer must be tested before first energization at site. The minimum tests include:
- Visual inspection — check for transport damage, oil leaks, loose connections
- Oil sample — test dielectric strength, moisture content, DGA baseline
- Insulation resistance — all winding combinations to ground and to each other
- Winding resistance — verify all tap positions
- Turns ratio — verify all tap positions
- Auxiliary wiring IR test — 1 kV DC, ≥1 MΩ per IEC 60076-3 Clause 9
- Protection system check — verify CT ratios, polarity, and relay settings
After transport or long storage
If the transformer has been stored or transported, moisture may have entered the insulation. Test the IR and compare to factory values. If the IR has dropped significantly, the transformer may need to be dried out before energization.
Drying indicators: IR below 70% of factory value, moisture content above 2% in paper, or oil dielectric strength below 40 kV/2.5mm (IEC 60156).
Setting Up a Transformer Testing Program
Recommended test intervals
| Test | Frequency |
|---|---|
| DGA (oil sample) | Every 6–12 months (annually minimum) |
| Oil quality (moisture, dielectric strength) | Every 1–2 years |
| Insulation resistance (IR) | Every 2–3 years or during outages |
| Tan delta / power factor | Every 3–5 years or during major outages |
| Partial discharge (online monitoring) | Continuous where installed |
| Winding resistance | During major outages |
| Turns ratio | During major outages |
| Full dielectric tests | Only after major repair or rewind |
DGA is the highest priority. If you can only afford one test, make it DGA. It catches the widest range of developing faults with the least effort.
Trending and data management
Record every test with: transformer ID, date, oil temperature, ambient temperature, tap position, test method, and all readings. Plot trends annually. Look for:
- Declining IR over consecutive tests
- Rising tan delta
- Increasing gas levels (especially CO, CO₂, acetylene)
- Any sudden changes from previous readings
Quick Reference: All Test Values in One Table
| Test | Standard | Test Voltage | Acceptance Criteria |
|---|---|---|---|
| IR (oil-filled transformer) | IEEE C57.152 | 1–10 kV DC | Trend-based; new >500 MΩ |
| IR (HV machinery) | IEC 60204-11, Cl. 19.3 | Rated V or 5 kV (lower) | ≥ 1 MΩ |
| PI (transformer) | IEEE C57.152, Cl. 7.2.13.4 | Same as IR | >2.0 good; <1.0 dangerous |
| Tan delta | Field practice | AC test voltage | <0.5% new; <2% in-service |
| PD during IVPD | IEC 60076-3, Cl. 11.3.5 | Induced voltage | ≤250 pC; ≤100 pC at 1.2Ur/√3 |
| Applied voltage (AV) | IEC 60076-3, Table 2 | Per Um (3 kV–630 kV) | No voltage collapse in 60s |
| Lightning impulse (LI) | IEC 60076-3, Table 2 | Per Um (20 kV–2250 kV peak) | No waveform deviation |
| Auxiliary wiring | IEC 60076-3, Cl. 9 | 2 kV AC | No breakdown in 1 min |
| CT secondary wiring | IEC 60076-3, Cl. 9 | 2.5 kV AC (4 kV if Vk >2 kV) | No breakdown in 1 min |
| Site wiring check | IEC 60076-3, Cl. 9 (note) | 1 kV DC | ≥ 1 MΩ |
| CT primary to ground | Field practice (HV CTs) | 5,000V DC | >100 MΩ |
| CT primary to secondary | Field practice | 2,500V DC | >100 MΩ |
| CT secondary to ground | Field practice | 1,000V DC | >100 MΩ |
| CT tan delta | Field practice (HV) | AC test voltage | <0.5% |
FAQ
How often should I test transformer insulation?
DGA every 6–12 months. IR testing every 2–3 years or during planned outages. Tan delta during major maintenance outages (every 3–5 years). More frequently for critical transformers, aging units, or those in harsh environments.
What’s the most important test for transformer insulation?
DGA. It detects the widest range of developing faults — thermal, electrical, and chemical — and catches problems that electrical tests miss. If your budget allows only one test, choose DGA.
Can I use IEEE 43 PI values for transformers?
No. IEEE 43-2013 applies to rotating machinery only. For transformers, use IEEE C57.152-2013, which has different (lower) PI thresholds. A PI of 1.5 is below minimum for a Class F motor but is “fair” for a transformer.
What does a low IR reading mean on a new transformer?
If the transformer was recently shipped, moisture may have entered during transport or storage. Retest after a drying cycle. If the oil was recently replaced, new oil can temporarily lower IR readings. Test again after the oil has been in service for a few weeks.
Do I need to test auxiliary wiring separately from the main windings?
Yes. Per IEC 60076-3-2013 Clause 9, auxiliary wiring is tested separately at 2 kV AC (control wiring) or 2.5 kV AC (CT secondary wiring). All electronic devices must be disconnected first. On site before energization, the standard practice is a 1 kV DC test with ≥1 MΩ minimum.
What’s the difference between routine, type, and special tests?
Routine tests are done on every transformer manufactured. Type tests are done on a representative unit to prove the design. Special tests are done only when specified by the purchaser at the time of order. Per IEC 60076-3 Table 1, PD measurement is a routine test for transformers above 72.5 kV.
Key Takeaways
- Over 60% of transformer failures involve insulation breakdown. Testing catches problems early.
- DGA is the most important single test for oil-filled transformers. Do it annually minimum.
- Transformer PI values are interpreted differently from motor PI values — use IEEE C57.152, not IEEE 43.
- IEC 60076-3-2013 Table 2 defines the factory dielectric test voltages by Um class.
- PD measurement (≤250 pC acceptance) is a routine test for all transformers above 72.5 kV.
- Auxiliary and CT wiring has its own test requirements: 2 kV AC for control, 2.5 kV AC for CT circuits.
- Always check auxiliary wiring at 1 kV DC (≥1 MΩ) on site before first energization.
- Trending is everything. A single reading is a snapshot. Years of data predict failures.
Standards Referenced in This Article
| Standard | Edition | Key Content Used |
|---|---|---|
| IEC 60076-3 | Ed. 3.0 (2013) | Table 1 (test requirements), Table 2 (test voltages), Clause 8 (after service), Clause 9 (auxiliary wiring), Clause 11.3.5 (PD acceptance) |
| IEC 60204-11 | Ed. 1.0 (2000) | Clause 19.3 (HV machinery IR: rated V or 5 kV, ≥1 MΩ) |
| IEEE C57.152 | 2013 | Clause 7.2.13.4 (transformer PI interpretation table) |
| IEEE 43-2013 | 2013 | Referenced for comparison — does NOT apply to transformers |
| IEC 60076-5 | Ed. 3.0 (2006) | Short circuit withstand (referenced for mechanical stress context) |